Early assays of plasma testosterone involved chromographic purification of plasma prior to RIA. A variety of commercial kits are available providing a direct assay using unextracted plasma. We compared the testosterone concentration via RIA both before and after extraction and purification on Sephadex LH-20 columns. Plasma samples were obtained from 36 males and female infants. Results of the RIA using unextracted plasma were significantly higher than those using plasma following purification. This discrepancy was most pronounced during the first 3 weeks of life. These data indicate that direct assays of plasma testosterone are inappropriate during the first month of life.